Examples of org.apache.hadoop.hbase.io.hfile.HFileScanner

Package org.apache.hadoop.hbase.io.hfile

Examples of org.apache.hadoop.hbase.io.hfile.HFileScanner

org.apache.hadoop.hbase.io.hfile.HFileScanner
A scanner allows you to position yourself within a HFile and scan through it. It allows you to reposition yourself as well.
A scanner doesn't always have a key/value that it is pointing to when it is first created and before {@link #seekTo()}/ {@link #seekTo(byte[])} are called.In this case, {@link #getKey()}/ {@link #getValue()} returns null. At mostother times, a key and value will be available. The general pattern is that you position the Scanner using the seekTo variants and then getKey and getValue.


  private void verifyCounts(int countRow1, int countRow2) throws Exception {
    int count1 = 0;
    int count2 = 0;
    for (StoreFile f: this.r.stores.get(COLUMN_FAMILY_TEXT).getStorefiles()) {
      HFileScanner scanner = f.getReader().getScanner(false, false);
      scanner.seekTo();
      do {
        byte [] row = scanner.getKeyValue().getRow();
        if (Bytes.equals(row, STARTROW)) {
          count1++;
        } else if(Bytes.equals(row, secondRowBytes)) {
          count2++;
        }
      } while(scanner.next());
    }
    assertEquals(countRow1,count1);
    assertEquals(countRow2,count2);
  }

View Full Code Here

      // If the row we're looking for is past the end of file, set search key to
      // last key. TODO: Cache last and first key rather than make each time.
      firstOnRow = new KeyValue(lastKV.getRow(), HConstants.LATEST_TIMESTAMP);
    }
    // Get a scanner that caches blocks and that uses pread.
    HFileScanner scanner = r.getScanner(true, true, false);
    // Seek scanner.  If can't seek it, return.
    if (!seekToScanner(scanner, firstOnRow, firstKV)) return;
    // If we found candidate on firstOnRow, just return. THIS WILL NEVER HAPPEN!
    // Unlikely that there'll be an instance of actual first row in table.
    if (walkForwardInSingleRow(scanner, firstOnRow, state)) return;
    // If here, need to start backing up.
    while (scanner.seekBefore(firstOnRow.getBuffer(), firstOnRow.getKeyOffset(),
       firstOnRow.getKeyLength())) {
      KeyValue kv = scanner.getKeyValue();
      if (!state.isTargetTable(kv)) break;
      if (!state.isBetterCandidate(kv)) break;
      // Make new first on row.
      firstOnRow = new KeyValue(kv.getRow(), HConstants.LATEST_TIMESTAMP);
      // Seek scanner.  If can't seek it, break.

View Full Code Here


  private int count() throws IOException {
    int count = 0;
    for (StoreFile f: this.r.stores.
        get(COLUMN_FAMILY_TEXT).getStorefiles()) {
      HFileScanner scanner = f.getReader().getScanner(false, false);
      if (!scanner.seekTo()) {
        continue;
      }
      do {
        count++;
      } while(scanner.next());
    }
    return count;
  }

View Full Code Here

      BloomType bloomFilterType = familyDescriptor.getBloomFilterType();
      
      halfWriter = new StoreFile.Writer(
          fs, outFile, blocksize, compression, conf, KeyValue.COMPARATOR,
          bloomFilterType, 0);
      HFileScanner scanner = halfReader.getScanner(false, false);
      scanner.seekTo();
      do {
        KeyValue kv = scanner.getKeyValue();
        halfWriter.append(kv);
      } while (scanner.next());
      
      for (Map.Entry<byte[],byte[]> entry : fileInfo.entrySet()) {
        if (shouldCopyHFileMetaKey(entry.getKey())) {
          halfWriter.appendFileInfo(entry.getKey(), entry.getValue());
        }

View Full Code Here

        writer, Bytes.toBytes("cf"), Bytes.toBytes("qual"));
    
    StoreFile sf = new StoreFile(fs, writer.getPath(), false,
        util.getConfiguration(), StoreFile.BloomType.NONE, false);
    HFile.Reader reader = sf.createReader();
    HFileScanner scanner = reader.getScanner(false, true);
    
    FaultyInputStream inStream = fs.inStreams.get(0).get();
    assertNotNull(inStream);
    
    scanner.seekTo();
    // Do at least one successful read
    assertTrue(scanner.next());
    
    inStream.startFaults();
    
    try {
      int scanned=0;
      while (scanner.next()) {
        scanned++;
      }
      fail("Scanner didn't throw after faults injected");
    } catch (IOException ioe) {
      LOG.info("Got expected exception", ioe);

View Full Code Here

  private int verifyHFile(Path p) throws IOException {
    Configuration conf = util.getConfiguration();
    HFile.Reader reader = new HFile.Reader(
        p.getFileSystem(conf), p, null, false);
    reader.loadFileInfo();
    HFileScanner scanner = reader.getScanner(false, false);
    scanner.seekTo();
    int count = 0;
    do {
      count++;
    } while (scanner.next());
    assertTrue(count > 0);
    return count;
  }

View Full Code Here

    return this.top;
  }


  @Override
  public HFileScanner getScanner(final boolean cacheBlocks, final boolean pread) {
    final HFileScanner s = super.getScanner(cacheBlocks, pread);
    return new HFileScanner() {
      final HFileScanner delegate = s;
      public boolean atEnd = false;


      public ByteBuffer getKey() {
        if (atEnd) return null;

View Full Code Here

  public byte[] getLastKey() {
    if (top) {
      return super.getLastKey();
    }
    // Get a scanner that caches the block and that uses pread.
    HFileScanner scanner = getScanner(true, true);
    try {
      if (scanner.seekBefore(this.splitkey)) {
        return Bytes.toBytes(scanner.getKey());
      }
    } catch (IOException e) {
      LOG.warn("Failed seekBefore " + Bytes.toString(this.splitkey), e);
    }
    return null;

View Full Code Here

      // If the row we're looking for is past the end of file, set search key to
      // last key. TODO: Cache last and first key rather than make each time.
      firstOnRow = new KeyValue(lastKV.getRow(), HConstants.LATEST_TIMESTAMP);
    }
    // Get a scanner that caches blocks and that uses pread.
    HFileScanner scanner = r.getScanner(true, true);
    // Seek scanner.  If can't seek it, return.
    if (!seekToScanner(scanner, firstOnRow, firstKV)) return;
    // If we found candidate on firstOnRow, just return. THIS WILL NEVER HAPPEN!
    // Unlikely that there'll be an instance of actual first row in table.
    if (walkForwardInSingleRow(scanner, firstOnRow, state)) return;
    // If here, need to start backing up.
    while (scanner.seekBefore(firstOnRow.getBuffer(), firstOnRow.getKeyOffset(),
       firstOnRow.getKeyLength())) {
      KeyValue kv = scanner.getKeyValue();
      if (!state.isTargetTable(kv)) break;
      if (!state.isBetterCandidate(kv)) break;
      // Make new first on row.
      firstOnRow = new KeyValue(kv.getRow(), HConstants.LATEST_TIMESTAMP);
      // Seek scanner.  If can't seek it, break.

View Full Code Here

    Path refPath = StoreFile.split(fs, dir, hsf, midRow, Range.top);
    StoreFile refHsf = new StoreFile(this.fs, refPath, true, conf, 
        StoreFile.BloomType.NONE, false);
    // Now confirm that I can read from the reference and that it only gets
    // keys from top half of the file.
    HFileScanner s = refHsf.createReader().getScanner(false, false);
    for(boolean first = true; (!s.isSeeked() && s.seekTo()) || s.next();) {
      ByteBuffer bb = s.getKey();
      kv = KeyValue.createKeyValueFromKey(bb);
      if (first) {
        assertTrue(Bytes.equals(kv.getRow(), midRow));
        first = false;
      }

View Full Code Here

0 1 2 3 4 5 6 7 8 9

TOP

Related Classes of org.apache.hadoop.hbase.io.hfile.HFileScanner

org.apache.crunch.io.hbase.HFileReaderFactory

org.apache.hadoop.hbase.HFilePerformanceEvaluation$GaussianRandomReadBenchmark

org.apache.hadoop.hbase.HFilePerformanceEvaluation$UniformRandomReadBenchmark

org.apache.hadoop.hbase.HFilePerformanceEvaluation$UniformRandomSmallScan

org.apache.hadoop.hbase.index.io.IndexHalfStoreFileReader

org.apache.hadoop.hbase.index.io.TestIndexHalfStoreFileReader

org.apache.hadoop.hbase.index.mapreduce.IndexLoadIncrementalHFile

org.apache.hadoop.hbase.io.HalfHFileReader

org.apache.hadoop.hbase.io.HalfStoreFileReader

org.apache.hadoop.hbase.io.TestHalfStoreFileReader

All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.